ClusterLabs/pacemaker Pacemaker-1.1.23

Pacemaker 1.1.23 - Final

on GitHub

Changesets: 115

Diff: 79 files changed, 2818 insertions(+), 1293 deletions(-)

Features added since Pacemaker-1.1.22

• fencing: improve parameter handling for RHCS-style fence agents
  - add port or plug parameter according to metadata
  - plug parameter shown as non-required in the metadata
• tools: add stonith_admin --delay option (default = 0) to support enforced
  fencing delay

Changes since Pacemaker-1.1.22

• crmd: properly detect node info requests from remote nodes
  (regression introduced in 1.1.19)
• crmd: remove node as voter only if lost
  (regression introduced in 1.1.20)
• crmd: clear leaving node's transient attributes even if there is no DC
• crmd: don't ack same request twice
• crmd: avoid delay when recovering resources after shutdown-lock expiration
  (BZ#5424)
• fencing: remove dangling 'pending' for completed DC fencing (BZ#5401)
• fencing: avoid possible use-of-NULL when parsing metadata
• fencing: "action" now shown as not required in agent metadata
• libcrmcommon: fix parsing ISO 8601 offsets with a plus sign
• libcrmcluster: use proper type for corosync id when updating node state
• pengine: don't incorporate dependencies' colocations multiple times
• pengine: ensure attenuated scores still have some effect
• pengine: ignore colocation constraints with 0 score
• pengine: make sure cluster-wide maintenance-mode=true overrides per-resource settings